Classifying pediatric brain tumors by liquid biopsy using AI (www.stjude.org)

Scientists at St. Jude Children’s Research Hospital, in collaboration with international colleagues, have developed an innovative AI-based tool called the Methylation-based Predictive Algorithm for CNS Tumors (M-PACT) for classifying pediatric brain tumors through liquid biopsies. This noninvasive method analyzes circulating tumor DNA (ctDNA) obtained from cerebrospinal fluid, overcoming previous technological limitations in detecting small ctDNA amounts in children. Published in *Nature Cancer*, M-PACT demonstrated a remarkable 92% accuracy in identifying brain tumors, differentiating between relapses and secondary tumors, and tracking changes in tumor aggressiveness without additional input. The significance of M-PACT lies in its capacity not only to enhance diagnostic accuracy in pediatric neuro-oncology but also to provide insights into the tumor microenvironment by analyzing noncancerous cell types that influence tumor behavior. This methodology represents a leap forward in methylation-based diagnostics, creating a new standard that may extend beyond brain tumors to various cancer types. By leveraging a large dataset of DNA methylation profiles, M-PACT paves the way for future applications in cancer monitoring and treatment, signaling a transformative shift in noninvasive cancer detection methodologies that could reshape clinical practices in oncology.